The photomodel comes without a Voltage regulator. Supply was 9 V, 200 mA. The final Version also has an Attenuator at the output. For Fine Adjustement of the Level and Improvement of the Matching. BOM is approx. 99 CHF.

We have heard, that there are budget friendly Noise Sources out there, but those I saw don't show a flat amplitude response. Yes, you can calibrate that away, but thats not what we wanted. So we searched a wideband transistor with a high Noise Figure. And found many of them. We now can tailor the ENR value just by changing the assembly. Nice.

Some theroretical max. ENR values (without attenuator) mayst be :

TYPE	NF [dB]	GAIN [dB]	ENR [dB]
GVA-81+	7.4	10	24.8
GVA-82+	6.6	13.8	27.0
GVA-83+	6.2	17.1	29.5

In order to calculate the ENR value from the Noise Floor Measurement, we used this website. The above ENR was measured with two identical Gain Blocks. The 50Ω Resistor produces a Noise of -113.1 dBm in a 300 kHz Bandwidth at 290 K.

With 2 x GVA-81+ and PAT-3 the ENR value is : 25.85 dB @ 1.5 GHz
And yes, we know that graphs end, where data becomes unflattering. Please believe me, I only have a FSP-3 from Rhode & Schwarz (3 GHz) High Precision Spectrum Analyzer 🙁

As some Tester will switch the Noise Source ON and OFF, the Output matching must be very good, in both cases. This can be accomplished by increasing the attenuator at the output to higher values. Using an attenuator of 10 dB, we can increase the matching by 20 dB. Yes, the ENR will drop by 10 dB, but that's not a problem. We measure the following values :

FREQUENCY [MHz]	ENR [dB]
500	21.60
1000	21.43
2000	20.90
3000	21.34

The webmaster does not read these comments regularely. Urgent questions should be send via email. Ads or links to completely uncorrelated things will be removed.